|Publication number||US8006186 B2|
|Application number||US 10/451,041|
|Publication date||Aug 23, 2011|
|Filing date||Dec 22, 2000|
|Priority date||Dec 22, 2000|
|Also published as||EP1354318A1, US20040027369, WO2002052565A1|
|Publication number||10451041, 451041, PCT/2000/197, PCT/SG/0/000197, PCT/SG/0/00197, PCT/SG/2000/000197, PCT/SG/2000/00197, PCT/SG0/000197, PCT/SG0/00197, PCT/SG0000197, PCT/SG000197, PCT/SG2000/000197, PCT/SG2000/00197, PCT/SG2000000197, PCT/SG200000197, US 8006186 B2, US 8006186B2, US-B2-8006186, US8006186 B2, US8006186B2|
|Inventors||Peter Rowan Kellock, Edward James Altman|
|Original Assignee||Muvee Technologies Pte. Ltd.|
|Export Citation||BiBTeX, EndNote, RefMan|
|Patent Citations (107), Non-Patent Citations (13), Referenced by (39), Classifications (12), Legal Events (2)|
|External Links: USPTO, USPTO Assignment, Espacenet|
The invention relates generally to computer generation of media productions. In particular, the invention relates to automated or semi-automated editing of productions including any one or more of the following media: motion video, still images, music, speech, sound effects, animated graphics and text.
Today, analogue media are gradually being replaced by digital media. In the case of audio this transition has already largely taken place, and it is well underway for image, video, graphics animation and other media. As these media become digital and the capability/cost ratio of computing resources continues to increase, new users and markets for digital media production are opening up. Of particular relevance to this invention are emerging markets for casual media production and especially casual video production, i.e. cases where the high cost of professional video production would preclude its use and where, until recently, the cost of the necessary equipment was too high. These include home video production (e.g. of holidays, weddings, etc), some informal corporate uses (e.g. internal communications and team-building), use by societies and other organizations, etc.
The concept of casual or “desktop” video production has existed for about a decade, but widespread adoption has been held back by a number of problems. These include:
The first and second of these problems are today disappearing thanks to technologies such as DV cameras, the IEEE 1394 (“Firewire”) interface and video distribution on the world-wide web.
This invention attempts to address the third problem, allowing automated or semi-automated editing of digital media, particularly video.
Today, the main tool used for editing video is the “Non-Linear video Editor” or NLE. These are computer programs which adopt many paradigms from conventional editing methods such as film cutting and linear dub-editing using video tape machines. They employ manual methods of editing which are well suited to scenarios where the user is experienced and the desired result is a high-quality video production. There are many products of this type including Premiere from Adobe Inc., and iMovie from Apple Inc.
The NLE is a considerable advance on earlier technology, yet there remain many scenarios in which the user is not a media professional, in which professional quality is not essential, or in which it is necessary to edit material very quickly. Even NLEs which claim to be aimed at non-professionals have a significant learning curve and require substantial time to produce acceptable productions. It is generally accepted that in typical cases a user will have to spend one hour in order to create one minute of output video, in other words a ratio of 60:1 of production time to playback duration. It is one of the goals of the current invention to reduce this ratio dramatically through automation, to the point where in some cases acceptable results can be produced without any user intervention.
There also exist several tools which allow a user to create productions involving the real-time display of images and text synchronized to an audio track. These include animation tools (e.g. Flash from Macromedia Inc.), slideshow tools (e.g. PowerPoint from Microsoft Inc.) and authoring tools for streaming media (e.g. RealPlayer from Real Networks Inc.). But once again, users often find that they need to spend hours in order to produce a simple production lasting a few minutes.
This invention aims to provide new and useful apparatus and methods for generating media productions from input media material.
In general terms, the invention proposes that input material is edited to construct an output production. The process includes deriving a set of media descriptors which describe the input material, either by analysis or from an external source, or a combination of the two. This is followed by a computer-based construction process which includes making edit decisions based on (i) the set of media descriptors, and (ii) style data, such as user-generated style data, for defining the editing style.
The input material may include any one or more of motion video, still images, music, speech, sound effects, animated graphics and text.
The set of media descriptors may be supplemented by descriptors which are pre-generated (e.g. outside the apparatus of the invention) and imported, for example together with the input material.
The style data may be generated by a process which includes either or both of deterministic and stochastic (probabilistic) operations.
The editing may include any one or more of the following processes applied to the input material: segmentation (of video/audio), selective inclusion, sequencing, transformation and combination. These processes may optionally be supplemented with user intervention. This is supported at two stages: a pre-selection stage prior to the automatic construction process and a touch-up stage after construction.
A particularly preferred feature of the invention is the ability to produce music-based productions in which the input material consists of a) motion video material and/or a set of images, and b) recorded music. The system analyses both the video/images and the music to create media description data for both, then uses this information to create the output production, influenced or determined by the structure of the music.
Typical applications of the current invention include the production of video and other time-based media for home, corporate and hobbyist environments, production of slideshows synchronized to music, production of rich-media electronic greeting cards, production of media for world-wide-websites, production of rich-media online catalogues, production of rich-media online content related to consumer-to-consumer sales applications such as online auctions and classified advertisements, some professional video applications such as the production of karaoke videos, etc.
The invention, which includes both method and apparatus aspects (i.e. apparatus comprising respective means to perform the steps of the methods), may be embodied within various kinds of hardware including general-purpose computers, personal digital assistants, dedicated video-editing boxes, set-top boxes, digital video recorders, televisions, games consoles, digital still cameras, and digital video cameras.
Embodiments of the invention are described hereinafter, for the sake of example only, with reference to the drawings, in which:
In this document the term “input material” is used to mean one or more pieces of media which are presented as input to the system. Supported media types include motion video, still images, music, speech, sound effects, static or animated graphics and static or animated text. The term “input visual material” refers to input material of any visual type including video, images, animation, graphics or text.
The output production may also include material generated by the system without reference to the input material, such as colors and textures used as backgrounds, static and animated graphical elements, etc.
Analysis and Description Components
Referring again to
In this document, the analyzer components described above [1110, 112, 114 and 116] are collectively known as the “media analyzers”, and the descriptions [111, 113, 115 and 117] are known as “media descriptions”.
Media descriptions can also be stored for later use, for example by saving the description data to disk or non-volatile memory. (For simplicity, this is not shown in
In addition to, or alternatively to, signal analysis, descriptors may be imported into the system and stored in the media descriptions. (For simplicity, this is not shown in
Imported descriptors can also originate from a manual or semi-automatic process, for example in which a user annotates the input music, video or images before importing the input material plus its descriptions into the system. Such descriptors may be closely related to descriptors created by signal analysis. For example it is sometimes desirable to create descriptors using the system, correct or refine these descriptions manually, and then use the refined description as the basis for processing by the other modules of the system.
Imported descriptors may be stored directly in the media descriptions, or they may require further analysis, conversion or interpretation after they are imported; this function is also provided by the media analyzers.
Referring again to
Features and variants of the GUI are further described below.
The MSG can optionally be saved and reloaded for later use, allowing progressive touch-up of the final production. Also, parts of the MSG (for example temporal regions or certain types of edit information) can be “locked” and others “unlocked”. This allows an output production to be made by progressive refinement: the user instructs the system to run the constructor (and renderer), views the resulting output production, locks regions or features that he/she likes, runs the constructor (and renderer) again to replace the unlocked regions/features, views the altered output production, locks another set of regions/features, and so on.
The logic of the constructor and the structure of the MSG are described in detail below.
In general in this document, the invention is described as a single system including the media analyzers, the constructor and the renderer. However it can also be a distributed system in which each of these modules is a separate program, potentially run at different times at different locations by different parties. It has already been mentioned that media descriptions can be stored and imported when needed by the constructor. Such media descriptions can be created by media analyzer modules invoked at any earlier time at any location by any party.
Likewise, because the MSG is a complete representation of the form of the output production or a complete set of instructions for making the output production, the renderer can be run separately from the constructor or analyzers. It can even be run in real-time while the output production is viewed, in other words creating the output production on the fly, in which case the renderer is in effect a sophisticated playback engine. All that is required to make this possible is that the MSG and the input material are available at the time of rendering.
For example, in an application where two parties share access to a common body of input material, or have two identical copies of the input material, one party can run the analyzers and constructor in order to create an MSG, then send this MSG to the second party whereupon the second party runs the renderer to create the output production “on the fly” as she/he views it. In another example, a community of people can first acquire copies of a common body of input material and associated pre-created media descriptions, then individually produce output productions which they share with each other simply by transmitting different MSG's. The advantage of this is that each MSG is a small amount of data compared to typical media data and can therefore be transmitted quickly and easily. The common body of media is suited to distribution on a medium such as CD-ROM or DVD; the community of people owning the CD-ROM/DVD can share their productions by, for example, forwarding MSG's as email attachments.
The process of automatic construction will now be described in detail with reference to
Video Editing Example
All the above involve timing references relative to the output production; these are shown as vertical dotted lines projected onto the timeline  of the output production. Segments of input video involve an additional set of timing references relative to their input video source file, for example the start time  and end time  of sSeg4.
In conventional NLEs, the user makes all decisions about which of these processes to apply and where to apply them. The current invention creates an output production automatically by making the decisions itself and invoking processes such as those above accordingly. The constructor  is the heart of the system and decides which processes to apply and where to apply them, while the renderer  performs the actual processing.
The Construction Process
Some examples of this process will now be presented in more detail, starting from the nature of styles.
Styles may be defined by data or logic or some mix of the two. For example, the style information  of
Style information is created by a style designer, for example by a process of manually defining a set of values for parameters, and the aim of the style designer is to create styles which will cause the system to generate high-quality output productions. The information comprising styles may be categorized according to which part of the construction process they affect, using a similar categorization to that used above. For example, the style information of one embodiment has the following:
The precise choice of parameters and their values is both highly dependent on context and partially subjective. The range of possibilities is enormous and influenced by factors such as the type and range of input material which must be handled successfully, the demographics and preferences of target users of the system, and other such factors.
Generating Variety in Edit Decisions
In order to create an interesting production, it is usually necessary to introduce some variation in the edit decisions through the course of a production. For example, in most cases it is desirable to vary the preferred segment duration introduced above. A production consisting of segments of identical length would quickly become tedious, so the duration of segments must be varied to create a satisfying “edit rhythm”.
In one embodiment, this need to introduce variety is addressed in several ways which may be used singly or in combination:
TABLE 1 Cycle Length 1 2 3 4 5 6 7 8 9 10 11 12 Target 2 Dark Bright Dark Bright Dark Bright Dark Bright Dark Bright Dark Bright Brightness Duration (s) 3 4 2 2 4 2 2 4 2 2 4 2 2 Transition 3 Cut Cut Diss. Cut Cut Diss. Cut Cut Diss. Cut Cut Diss. type Color or 4 M C C C M C C C M C C C Monochrome? Flash effect? 8 No No No No No No No Yes No No No No Time → → → → →
Selection of Elements of the Input Material to Construct the Output Production
A central function of the constructor is to select and sequence the elements of the input material (such as segments of video, music or soundtrack, selected images, or regions within images or video frames) which will form the output production. This will now be described for cases where the input material is video and the elements of the input material in question are video segments. The process for other media such as a set of images is related and generally simpler.
As described above (see “sequencing rules”), styles specify whether the elements of the input material comprising the output production are to be chosen sequentially, non-sequentially or partly-sequentially from the input material. The process of selecting elements of the input material involves a number of complexities which will now be explained with reference to a sequential case and a non-sequential case. Variants of these cases, such as partly-sequential cases, can be achieved using a mix of the techniques described in the following.
Segment Selection: a Sequential Case
The constructor builds the output production segment by segment. In this example, it has already built 6 segments, O1 to O6 and is about to build the next segment. To select a new segment the constructor follows the flowchart in
The constructor first gets the start time in the output production  for the new segment , labeled to in
The constructor then calculates a time ti in the input video at which it will start looking for a suitable segment . In sequential cases it is generally desirable that the ouput production be approximately linear with respect to the input video, and to achieve this the input video segment should ideally be taken from a time location calculated as follows:
t i =R io *t o
In other words the relative position in the input and output videos should be the same.
The constructor checks whether there is a subsegment of the segment at ti which is long enough to form the new segment—i.e. which is at least dT in duration . In addition to having duration >=dT, the choice of subsegment is subject to two constraints:
If such a piece cannot be found within the input segment at ti, the constructor searches forward  into later segments looking for a segment which is long enough (duration >=dT). However there is no point searching too far forward: selecting a segment far ahead of the current location in the input video would not allow later segments to be sequential. A suitable location in the input video at which to stop searching is given by the formula ti-stop=Rio*(to+do).
If the constructor finds a segment or subsegment from the above, it then chooses a piece of duration dT from within it  to use as the output segment. The choice of this piece can be simple: it may, for example, simply choose the first part of the subsegment. Or it be sophisticated, attempting to find a piece of length dT which meets other criteria, for example by matching descriptor target values (using similar principles to those described below for a non-sequential case) or by selecting pieces which are estimated to be more interesting or superior in quality to the surrounding material (also see below); this is most useful when the segments of input video are significantly longer than the segments of the output video, a common situation.
If the constructor is unable to find a suitable segment from either of the above approaches, it relaxes the constraint that an output segment should not contain a segment boundary in the input video and builds an output segment of duration dT from two or more segments/subsegments of the input video .
Segment Selection: a Non-Sequential Case
In this non-sequential case (
As in the above, the Constructor first gets the start time in the output production for the new segment  and then gets a set of parameters required for the new segment, for example from the Style Information, including the required segment duration do and data about effects and transitions . In this non-sequential case, it also gets a set of target descriptor values from the style information ; it will select segments which match this set of values.
The duration then has to be adjusted to produce a target segment duration dT for the segment which is to be taken from the input video , allowing for transitions and speed changes in the same way as described above for a sequential case.
The next step  is to find a set of candidate segments or subsegments of the input video. These are segments which are at least dT in duration. They may also have to satisfy other criteria. For example, although some re-use of material may be permitted in certain non-sequential cases (unlike a strictly sequential case) it is generally desirable to limit the number of times the same material appears in the output production. This can be achieved by keeping a count of how often each part of the input video has been used, in which case a candidate (sub)segment is any contiguous part of a single segment of the input material that has been used less than the maximum permitted number of times and is at least dT in duration.
If no such (sub)segment can be found, the constructor relaxes a constraint—for example, as in the sequential case above, it may build an output segment of duration dT from two or more segments/subsegments of the input video. (Not shown in figure.)
The Constructor then gets descriptor values from the media description for these “candidate” (sub)segments  and evaluates a distance in descriptor space between each of these candidate points and the target descriptor values . (This process is further described below, and shown in expanded form in
Selecting Segments by Proximity in Descriptor Space
As mentioned above, there is a need to select a best-match (sub)segment from a set of candidate (sub)segments. The best match is the (sub)segment which lies closest to the set of target values in the “descriptor space” (an n-dimensional space in which each of n descriptors is represented)—i.e. for which a distance measure between the given point (coordinates defined by the target values from the style information) and the candidate point (coordinates defined by the set of values in the media description) is smallest.
Although simple in principle, there are several issues to consider in this matching process. They will be described with reference to
The above describes direct matching of descriptors in which the types of descriptor in the Style Information are identical to the types of descriptor in the media description: for example a brightness descriptor in the Style Information is matched to a brightness descriptor in the media description. It is also possible to use indirect matching, in which the set of descriptors used in the Style Information are mapped to a different set of descriptors in the media description via a mathematical or algorithmic relationship. For example the style information might have a “PeopleEnergy” descriptor defined as:
where “Activity” is a measure of the total average movement in a video segment and “PersonProbability” is a measure of the likelihood that the segment contains images of at least one person (for example using a skin-color detection algorithm well-known in the art). Such descriptors, defined by a mathematical or algorithmic manipulation applied to one or more other descriptors, may be termed “derived descriptors”.
It is sometimes effective to define target values in the style information in terms of such derived descriptors, because this allows the use of “higher-level” descriptors which are closer to the kind of semantically-meaningful descriptors understood easily by human beings. In the above example, the style information would contain target values for PeopleEnergy while “Activity” and “PersonProbability” would be extracted by signal analysis of the input video.
If derived descriptors are used, the constructor logic can calculate values for the derived descriptors by applying mathematical or algorithmic manipulation to the lower level descriptors of the media description and then perform descriptor matching on the values of the derived descriptors.
Optimizing the Selection of Material
As mentioned above, the amount of input material is often much greater than the length of the output production and it is therefore desirable in some cases to select parts of the input material which are estimated to be more interesting or superior in quality to the rest of the material This is related to segment selection as described above, and it may use some of the same techniques, but its purpose is somewhat different: segment selection as described above is primarily concerned with where parts of the input material should be placed in the output production, whereas optimizing the selection of material is primarily concerned with which parts of the input material should be used in the output production.
Present technology does not provide techniques to determine the semantic content of video or images reliably across all kinds of material using signal analysis techniques. It is therefore impossible for an automatic system to select material exactly as a human video editor would do it. Furthermore, this is a highly subjective subject: different human editors would select different material. Nevertheless it is possible to bias the selection of material in such a way that the majority of users will judge it to be more interesting or superior in quality to the average of the input material. To put it another way, the goal is automatically to select material which is generally “better”, at least within certain types of material, than an unbiased set of samples taken from the input material.
Here are two examples of how this may be achieved:
These techniques may be provided as options which a user of the system can invoke when desired. Alternatively they may be invoked when the user selects certain styles and not invoked for other styles: for example the low-motion handheld techniques just described may be desirable in a style intended to produce output productions which are soothing or “laid back”, but not suitable in a style intended to create high-energy, fast-paced productions.
The Media Scene Graph (MSG)
As explained above, the MSG is a data structure produced by the constructor which completely represents the form of the output production. In this sense it is related to the concept of an edit decision list (EDL) well known in the prior art. However the MSG is also the primary underlying data structure which the user interacts with during touch-up, while a typical EDL is a linear structure which does not lend itself well to this kind of manipulation. An MSG structure which is better-suited to this kind of manipulation will now be described with reference to
The structure is essentially a tree in which the output production is the root . Some of the branches of the tree are concerned with definitions; these specify the properties of certain entities which are used in the output production. They include a definition for every kind of transformation  used in the Output Production (e.g. specifying that a certain flash effect has a certain brightness, color, radius, duration, etc). They also include definitions for transitions  such as dissolves, definitions for text  including animated text, definitions for graphic elements  including animated graphics, definitions for blank backgrounds , etc.
The MSG also has one or more branches for timelines.
The entries for segments, overlays, etc which comprise the timelines contain time data relating both to the output production, and in some cases to the input material. For example, for video segments it is necessary to specify a location (such as a start-time) and a duration in the output production; it is also necessary to specify the source such as a start-time and duration in the input material.
Graphical User Interface (GUI)
Due to the high degree of automation in the production process, the invention can in some cases produce an output production of acceptable quality without human intervention. Thus, in certain embodiments of the invention, the GUI can be very simple, or indeed non-existent.
An example of a very simple, yet viable GUI is shown in
There is also a standard control for closing the program .
A variant of this is shown in
There is also a standard control for closing the program .
In either of the above examples the output production can be viewed from an external program such a media player. Alternatively, the GUI elements described above can be supplemented with a viewing window and “transport controls” well-known in the art, so that the user can view the output production from within the system.
In other embodiments, the GUI may include additional features for manual interaction. The motivation for these is that, although the primary purpose of the invention is to automate the editing process, it cannot always do this completely in every case. Depending on the nature of the input material and the application in question, an output production created fully automatically may not match the user's preferences in every detail. Thus it may be desirable to support features for manual interaction such as:
The GUI for supporting these manual operations can be constructed using standard elements including lists, hierarchical representations (such as those used in file managers), visual thumbnails, audio waveform displays, timelines, clip windows with transport controls, etc. These elements are well known in the art, being common in tools such as Non-Linear Video Editors (NLE's), image editors, audio editors and other media-processing software.
The invention can also be embodied in a non-interactive system which simply presents output productions and does not include any GUI elements for normal use (although such a system does require a GUI for configuring and managing it). Logic for an example of such an embodiment is illustrated in
The embodiment is particularly suited to creating output productions in which the processing and timing of visual elements is governed by the characteristics and timing of an underlying music track. This is sometimes called “cutting to music” and is common in music videos, animated productions, promotion and marketing videos, television commercials and many other forms. Such productions are referred to as “music-based productions” in this document.
The general principle of music-based productions is that the music acts as the time reference. The visual elements are manipulated to conform to the music, but the music itself is not altered. Visual elements to which this may apply include motion video, images, animation, graphics and text. In addition, some non-musical audio elements such as speech and sound effects may be manipulated or positioned in time in ways which are influenced by the music. In general terms, the music is “master” and the other elements are “slaved” to it.
Music-based productions are constructed using a number of techniques. These techniques, today achieved through the skill of professional editors, include the following:
In summary, when visual material is to be edited to match music, the professional editor has available a repertoire of techniques across a range of timescales, from the “microstructure” of musical beats or even subdivisions of beats, all the way up to the “macrostructure” of the main sections comprising the piece of music. When this is done successfully, the effect on the viewer/listener is enhanced: music and video are more likely to be perceived as a unified production and the emotional or dramatic impact is enhanced.
The embodiment automates the creation of music-based productions in several ways which will now be described.
Automation for Music-Based Productions
The nature of the music analyzer  and music description  have been presented above and we have already introduced several ways in which the creation of music-based productions can be automated or facilitated. This aspect of the invention will now be further described.
One simple way to match editing style to music structure is to control the editing parameters defining the visual character of the output production directly from parameters of the music description. For example, the tempo of the music can be used to control the cutting speed (the inverse of the average segment duration), beat-strength used to control the ratio of cuts to dissolves, and loudness used to control the brightness of segments selected from the input video. In a straightforward mapping of this kind, a fast-cut output production will result if the user selects a piece of music with a fast tempo. Or, to take another example, if the user selects a piece of music with contrasting loud and quiet sections, the output production may have corresponding bright and dark sections.
This approach is effective in some cases, and the invention allows for it to be supported: for example, it can be implemented in certain styles, so that the user can select this mode of operation by selecting those styles. However, this approach has limitations because it relinquishes nearly all control to the music. For example, if the music is very uniform, the output production may be monotonous, because the mechanisms described above for introducing variety are not active. Conversely, if the music has many rapid contrasts, the output production may lack coherency. So this approach tends to lack robustness to different pieces of music: it may produce acceptable output productions for some pieces of music, but is not guaranteed to work for a wide range of musical pieces.
A more sophisticated alternative is to select styles and/or sub-styles according to the characteristics of the music, but then to allow the style information to control or influence the individual edit decisions. This produces results which are more predictable and coherent for any input music, because all edit decisions may be placed within bounds allowed by the style information. It also allows the style information to create variety even when the music is very uniform, for example using the techniques of stochastic generation and value cycling described above.
This approach conforms more closely to the central construction principle of the invention described above with reference to
As in the previous case discussed with reference to
The music macro-description  contains a description of the input music at the timescale of major sections of the music, such as introduction, verse, chorus, etc. The characteristics of these sections are represented by a set of music section descriptors which are used to produce a sub-style sequence . As mentioned above, the sub-style sequence defines the order in which the sub-styles are to be used to generate the output production. Once the sub-style sequence has been established, there exists, for any time in the output production, a corresponding sub-style. Thus, when edit information is required for a particular time in the output production, that information will be supplied by the correct sub-style.
The role of the music micro-description  will now be described. Referring back to the case, described earlier, where there is no input music, the information passed from styles/sub-styles to the construction logic  is effectively a set of edit commands, and the construction logic attempts to obey these commands if at all possible. (It may not always be possible, as some decisions depend upon the video/images description—see the discussion above about video segment selection—but generally it is possible and where it is, the construction logic will obey the command.)
However, in the case of music-based productions the information which the sub-style passes to the construction logic is a set of preferences: these preferences are to be followed only after the local features of the music, derived from the music micro-description , are considered. The micro-description contains a description of the input music at the timescale of bars, beats and sub-beat. This description can include, or be used to generate, a series of “edit hints”. For example, one kind of edit hint, which can be derived directly from a music amplitude descriptor, indicates that it is desirable to produce a segment transition in the output production at a certain time such as on a strong beat of the music.
Once the sub-style sequence has been created, the construction logic  is able to build the MSG as follows, starting from the beginning of the output production and traversing to the end of the output production:
The two major aspects of the above will now be described in greater detail by example: first, how a sub-style sequence matched to music macro-structure can be created, and second, a way in which the constructor can combine edit preferences with edit hints to produce edit decisions.
Creating a Sub-Style Sequence Matched to Music Macro-Structure
The general principle used to create a sub-style sequence matched to the music macro-structure is to use descriptor matching, a similar technique to that described in detail above for selecting input video segments by descriptor matching.
The goal of this process is to produce a sub-style sequence linked to the music structure such as the example shown in
Before proceeding it is worth noting two features of this example. First, each time the same music or similar music occurs, it is linked to the same sub-style: for example the chorus is always linked to SS4 in this case. This is normally desirable whenever the music sections are very similar, and the procedure about to be described will cause this result in many such cases. Secondly, there is no requirement for all the sub-styles of a particular style to be used: there is no “SS1” in this figure, implying that sub-style 1 has not been selected for this particular piece of music.
The next step,  is to retrieve from the style information a set of target descriptor values, one set for each sub-style. The set of target values in a sub-style constitutes a description of the characteristics of music which this sub-style would be particularly well matched to. Typically these are created by the style designer by a manual process of entering a set of target values for each sub-style. For example, when the style designer creates a fast-cut sub-style (i.e. one which contains or generates small values for the preferred segment duration, introduced above), she/he might define that this sub-style is best suited to a music section which exhibits high values for the tempo and beat-strength descriptors, but is not dependent on loudness.
The next step,  is to calculate a set of distances in descriptor space between music sections and sub-styles. This is similar to the process described above for selecting input video segments in a non-sequential case, and the techniques introduced for optimizing the calculation of proximity (see
The next step  is to check the sub-style sequence for undesirable repeats. This is necessary because, even if techniques such as descriptor distribution normalization (described above in connection with
If such undesirable repeats are found, they are eliminated  by replacing some of the sub-styles in the sub-style sequence with alternatives such as the next-nearest sub-style for each music section found in step  above.
Because this technique is similar to the techniques for selecting input video segments described in above with reference to
Combining Edit Preferences with Edit Hints to Produce Edit Decisions
In this example, the horizontal axis is time, and the vertical arrows  are edit hint pulses received or derived from the music micro-description. The height of these arrows is related to a perceptually-significant characteristic of the music and their horizontal location indicates the time at which they occur relative to a start-time t=0. Typically the characteristic in question is one which is closely related to the musical beat, such as a signal derived from the amplitude variations in the audio signal. Many techniques are known in the art for automatically extracting such representations of musical beat: for example, the overall amplitude, or the amplitude of a frequency band within the signal, can be subjected to a threshold-crossing test. Further refinements, such as the use of a phase-locked loop, can synchronize the detection mechanism with the periodicities in amplitude variation which occur when the beat is regular, as it is in most popular music. Whatever technique is used, it is desirable that the edit hint pulses have the following tendencies:
In this case the construction logic will interpret each edit hint pulse as a request to perform a cut at the corresponding time, and the height of each pulse as the strength of the request. The pulse height can be limited to a range such as 0 to 1; this is the case in
However, the construction logic also has to take account of the style/sub-style information. One parameter specified by styles is the “cutting speed” as introduced earlier. What is relevant to this example is that the style information specifies, for any moment in the output production, a preferred segment duration for the next shot of the output production. This preferred duration is marked tpreferred in
Two more facts are required to fully understand the operation of this mechanism:
Taking into account all the above factors, it can be seen that this thresholding mechanism exhibits the following behaviors:
It is often effective to set the value of tpreferred to a duration which is related to the beat speed at the current music tempo, for example, 1 beat, ½ beat, 2 beats, etc. Note also that the constructor often assigns tpreferred a different value for each segment as it progresses through the output production, using techniques such as those described earlier for creating variety in edit decisions: the use of sub-styles, gradual evolution, stochastic generation and value cycling.
By combining the set of techniques described in this section, the invention is able to generate edit decisions which are perceived as relating to the rhythm of the music, which are sufficiently varied even if the music is very regular, and which always lie within acceptable limits, regardless of the selected music.
Other Features for Automating the Creation of Music-Based Productions
The invention may optionally be enhanced with several other features for automating or facilitating the creation of music-based productions, for example:
Beat strength and other descriptors derived by the music analyzer can be used to control the animation of text/graphic overlays. For example, parameters of an overlay such as its location, orientation, size, skewing, color, etc can be determined directly by the amplitude of the music signal. Or, in a more sophisticated implementation, representations of musical beat based on a threshold-crossing test (as introduced above) can be used to trigger sudden changes in parameters of an overlay, and the overlay then allowed to relax to its default position rather more slowly. In other words the animation can be based upon a relaxation model which is excited by pulses derived from the music signal and related to the musical beat. Furthermore, the music section descriptors introduced above can be used to control changes in the animation behavior which is aligned with the section boundaries and is related to the musical characteristics of each section; for example the color, size and relaxation speed of an overlaid text/graphic animated as above could be made proportional to the average loudness of the current music section, so that overlays occurring during loud music will be large, bright and move in a jerky manner, while overlays occurring during quiet music will be small, dark and move more flowingly.
Changes to the Production Workflow
This final section describes how a typical embodiment of the invention changes the workflow for a user engaged in creating a media production, with reference to
As this example concerns a music-based production, the user also has to capture/import music , for example by recording it, or transferring it from a musical medium such as an audio CD, or acquiring music as a digital audio file. In any of these alternative scenarios, when this step is complete, the input music has been introduced into the NLE.
Some NLE's are able to perform the next step  automatically, segmenting the input video into shots using techniques such as detecting sudden changes in color histogram. The shots are presented to the user, typically as a set of “clips”, i.e. small segments of input video. If the NLE does not include automatic shot segmentation, the user segments the input video manually.
Next the user needs to familiarize herself/himself with the shots of the input video. This is typically done by “logging” the shots —i.e. organizing them in groups or in certain orders, making notes about each shot, rejecting some shots, etc. For professional productions involving a lot of input material this is usually a lengthy task. For small casual productions it may largely be bypassed, although doing so is usually detrimental to the quality of the resulting production.
The next three steps [1605, 1606, 1607] may be performed sequentially, or the user may alternate between them (for example finishing one section of the output production before moving on to the next section) or the user may work in a way which blurs the distinction between them. Whichever approach he/she adopts, the user must build the output production manually segment by segment, and—if a stylish music-based production is the goal—must carefully manipulate the segments so that they conform to the rhythm, timing and “feel” of the input music. This involves many of the techniques described above and is very time-taking in most cases, often requiring an hour, or several hours, to create each minute of the output production. It is also beyond the skill of many non-professional users to create a output production to a quality standard that they are happy with, particularly in the case of music-based productions, which require an understanding of music as well as visual material.
When the user believes that he/she has arrived at a satisfactory set of edit decisions, he/she instructs the NLE to render , at which point it produces an output production as a video file or other output. The user views this and, if not satisfied , returns to one of the earlier steps to alter or refine the production.
Finally the user exports their output production in a form which allows them, and others to view it . In the most basic case they may simply use the video file on their computer for local viewing, but more commonly they will transfer it to tape using a video cassette recorder, or to an optical disk format such as writeable compact disc (CD-R). It is also becoming increasingly common to distribute the video file using the Internet, for example by sending it as an email attachment, uploading it to a server which others can access, or sharing it from the user's local machine using so-called “peer-to-peer” file sharing.
The capture/import steps [1701 and 1702] are the same as the corresponding steps [1601 and 1602] described above for the conventional NLE case. The shot segmentation step  is also essentially the same as the correspond step in the above . The system uses one or more known techniques to automate the segmentation, and may optionally allow the user to override or adjust the resulting segmentation.
Next the user pre-selects content (elements of the input material) and/or treatment of the material . The invention provides techniques for assisting this process as described earlier. This step is optional and may be bypassed in some embodiments.
The next step  comprises the many kinds of automatic analysis and construction which have been extensively described in this document. Once this step is finished, a complete set of edit decisions has been generated—these fully define an output production. Typically this step is performed fully automatically by the system and requires no help from the user.
The system now renders the output production . The user views this and, if not satisfied , may either touch up the production with assistance from the system based upon the techniques described earlier , or may return to any of the earlier steps.
Finally the user exports their output production . This step is similar to the corresponding step  described above for the conventional NLE case.
It can be seen from
It will be clear to those skilled in the art that the invention can be embodied in many kinds of hardware device, including general-purpose computers, personal digital assistants, dedicated video-editing boxes, set-top boxes, digital video recorders, televisions, computer games consoles, digital still cameras, digital video cameras and other devices capable of media processing. It can also be embodied as a system comprising multiple devices, in which different parts of its functionality are embedded within more than one hardware device.
Although the invention has been described above with reference to particular embodiments, various modifications are possible within the scope of the invention as will be clear to a skilled persion.
|Cited Patent||Filing date||Publication date||Applicant||Title|
|US5262865||Aug 4, 1992||Nov 16, 1993||Sony Electronics Inc.||Virtual control apparatus for automating video editing stations|
|US5267034||Feb 25, 1992||Nov 30, 1993||Institute For Personalized Information Environment||Camera work detecting method|
|US5286908||Apr 30, 1991||Feb 15, 1994||Stanley Jungleib||Multi-media system including bi-directional music-to-graphic display interface|
|US5303041||Nov 25, 1992||Apr 12, 1994||Sony Corporation||Special effects generator with key signal motion detection|
|US5309228||May 21, 1992||May 3, 1994||Fuji Photo Film Co., Ltd.||Method of extracting feature image data and method of extracting person's face data|
|US5339393 *||Apr 15, 1993||Aug 16, 1994||Sony Electronics, Inc.||Graphical user interface for displaying available source material for editing|
|US5537530||Jan 23, 1995||Jul 16, 1996||International Business Machines Corporation||Video editing by locating segment boundaries and reordering segment sequences|
|US5541662 *||Sep 30, 1994||Jul 30, 1996||Intel Corporation||Content programmer control of video and data display using associated data|
|US5623424 *||Jun 6, 1995||Apr 22, 1997||Kabushiki Kaisha Toshiba||Rate-controlled digital video editing method and system which controls bit allocation of a video encoder by varying quantization levels|
|US5635982 *||Jun 27, 1994||Jun 3, 1997||Zhang; Hong J.||System for automatic video segmentation and key frame extraction for video sequences having both sharp and gradual transitions|
|US5655117 *||Sep 24, 1996||Aug 5, 1997||Oracle Corporation||Method and apparatus for indexing multimedia information streams|
|US5684715 *||Jun 7, 1995||Nov 4, 1997||Canon Information Systems, Inc.||Interactive video system with dynamic video object descriptors|
|US5693902||Sep 22, 1995||Dec 2, 1997||Sonic Desktop Software||Audio block sequence compiler for generating prescribed duration audio sequences|
|US5758180 *||Apr 15, 1993||May 26, 1998||Sony Corporation||Block resizing function for multi-media editing which moves other blocks in response to the resize only as necessary|
|US5794178 *||Apr 12, 1996||Aug 11, 1998||Hnc Software, Inc.||Visualization of information using graphical representations of context vector based relationships and attributes|
|US5801685||Apr 8, 1996||Sep 1, 1998||Tektronix, Inc.||Automatic editing of recorded video elements sychronized with a script text read or displayed|
|US5805733 *||Dec 12, 1994||Sep 8, 1998||Apple Computer, Inc.||Method and system for detecting scenes and summarizing video sequences|
|US5826102 *||Sep 23, 1996||Oct 20, 1998||Bell Atlantic Network Services, Inc.||Network arrangement for development delivery and presentation of multimedia applications using timelines to integrate multimedia objects and program objects|
|US5852435 *||Apr 12, 1996||Dec 22, 1998||Avid Technology, Inc.||Digital multimedia editing and data management system|
|US5861880||Oct 10, 1995||Jan 19, 1999||Fuji Xerox Co., Ltd.||Editing system for multi-media documents with parallel and sequential data|
|US5877445||Oct 24, 1997||Mar 2, 1999||Sonic Desktop Software||System for generating prescribed duration audio and/or video sequences|
|US5937136 *||Mar 6, 1997||Aug 10, 1999||Olympus Optical Co., Ltd.||Video data edit/reproduction apparatus for video data which mixedly includes moving and still images|
|US5956453 *||Apr 9, 1997||Sep 21, 1999||Hitachi Denshi Kabushiki Kaisha||Method of editing moving image and apparatus of editing the same|
|US5969716 *||Aug 6, 1996||Oct 19, 1999||Interval Research Corporation||Time-based media processing system|
|US5973679 *||Mar 31, 1997||Oct 26, 1999||Silicon Graphics, Inc.||System and method for media stream indexing|
|US5982447 *||Dec 11, 1996||Nov 9, 1999||Sony Corporation||System and method for combining two data streams while maintaining the continuous phase throughout the combined data stream|
|US6026389||Aug 4, 1997||Feb 15, 2000||Kokusai, Denshin, Denwa, Kabushiki Kaisha||Video query and editing system|
|US6032156 *||Apr 1, 1998||Feb 29, 2000||Marcus; Dwight||System for automated generation of media|
|US6070167 *||Mar 2, 1998||May 30, 2000||Sharp Laboratories Of America, Inc.||Hierarchical method and system for object-based audiovisual descriptive tagging of images for information retrieval, editing, and manipulation|
|US6118444 *||Jul 10, 1996||Sep 12, 2000||Avid Technology, Inc.||Media composition system with enhanced user interface features|
|US6137483||Nov 27, 1996||Oct 24, 2000||Matsushita Electric Industrial Co., Ltd.||Multimedia editing method using templates describing common logic structure, layout, and physical operation|
|US6137945 *||Apr 9, 1997||Oct 24, 2000||Sony Corporation||Apparatus and method for editing recorded material|
|US6138130 *||Jun 15, 1998||Oct 24, 2000||Inventure Technologies, Inc.||System and method for processing data in an electronic spreadsheet in accordance with a data type|
|US6141001 *||Aug 20, 1997||Oct 31, 2000||Alcatel||Method of synchronizing the presentation of static and dynamic components of an interactive multimedia document|
|US6163510||Jun 30, 1998||Dec 19, 2000||International Business Machines Corporation||Multimedia search and indexing system and method of operation using audio cues with signal thresholds|
|US6173275 *||Sep 17, 1997||Jan 9, 2001||Hnc Software, Inc.||Representation and retrieval of images using context vectors derived from image information elements|
|US6195458 *||Jul 29, 1997||Feb 27, 2001||Eastman Kodak Company||Method for content-based temporal segmentation of video|
|US6204840 *||Apr 8, 1998||Mar 20, 2001||Mgi Software Corporation||Non-timeline, non-linear digital multimedia composition method and system|
|US6226608 *||Jan 28, 1999||May 1, 2001||Dolby Laboratories Licensing Corporation||Data framing for adaptive-block-length coding system|
|US6243087 *||Sep 28, 1999||Jun 5, 2001||Interval Research Corporation||Time-based media processing system|
|US6243713 *||Aug 24, 1998||Jun 5, 2001||Excalibur Technologies Corp.||Multimedia document retrieval by application of multimedia queries to a unified index of multimedia data for a plurality of multimedia data types|
|US6262723 *||Nov 27, 1998||Jul 17, 2001||Matsushita Electric Industrial Co., Ltd.||System for use in multimedia editor for displaying only available media resources to facilitate selection|
|US6263149 *||Jul 14, 1999||Jul 17, 2001||U.S. Philips Corporation||Editing of digital video information signals|
|US6292620 *||Aug 13, 1999||Sep 18, 2001||Sony Corporation||Edited-list creating apparatus, editing apparatus and editing method|
|US6310839||Oct 23, 2000||Oct 30, 2001||International Business Machines Corporation||Multimedia search and indexing for automatic selection of scenes in a media by analyzing frequency ranges|
|US6320598 *||Mar 28, 2001||Nov 20, 2001||Interval Research Corporation||Time-based media processing system|
|US6330004 *||Nov 27, 1998||Dec 11, 2001||Matsushita Electric Industrial Co., Ltd.||Multimedia program editing and presenting system with flexible layout capability by simplified input operations|
|US6353170||Sep 3, 1999||Mar 5, 2002||Interlego Ag||Method and system for composing electronic music and generating graphical information|
|US6360234 *||Aug 14, 1998||Mar 19, 2002||Virage, Inc.||Video cataloger system with synchronized encoders|
|US6363380 *||Jan 13, 1998||Mar 26, 2002||U.S. Philips Corporation||Multimedia computer system with story segmentation capability and operating program therefor including finite automation video parser|
|US6374260 *||Feb 28, 2000||Apr 16, 2002||Magnifi, Inc.||Method and apparatus for uploading, indexing, analyzing, and searching media content|
|US6377519||Oct 23, 2000||Apr 23, 2002||International Business Machines Corp.||Multimedia search and indexing for automatic selection of scenes and/or sounds recorded in a media for replay|
|US6392133 *||Oct 17, 2000||May 21, 2002||Dbtech Sarl||Automatic soundtrack generator|
|US6449422 *||Nov 21, 1997||Sep 10, 2002||Sony Corporation||Editor, editing system, and editing method|
|US6452612 *||Dec 18, 1998||Sep 17, 2002||Parkervision, Inc.||Real time video production system and method|
|US6452875 *||Oct 23, 2000||Sep 17, 2002||International Business Machines Corp.||Multimedia search and indexing for automatic selection of scenes and/or sounds recorded in a media for replay by setting audio clip levels for frequency ranges of interest in the media|
|US6457027 *||Apr 1, 1997||Sep 24, 2002||Ati Technologies, Inc.||Method and apparatus for generating a multimedia document|
|US6463444 *||Aug 14, 1998||Oct 8, 2002||Virage, Inc.||Video cataloger system with extensibility|
|US6466239 *||Jan 26, 1998||Oct 15, 2002||Sony Corporation||Method and apparatus for editing data used in creating a three-dimensional virtual reality environment|
|US6469749 *||Oct 13, 1999||Oct 22, 2002||Koninklijke Philips Electronics N.V.||Automatic signature-based spotting, learning and extracting of commercials and other video content|
|US6490370 *||Aug 10, 1999||Dec 3, 2002||Koninklijke Philips Electronics N.V.||System and method for describing multimedia content|
|US6492998 *||Jan 29, 1999||Dec 10, 2002||Lg Electronics Inc.||Contents-based video story browsing system|
|US6542869||May 11, 2000||Apr 1, 2003||Fuji Xerox Co., Ltd.||Method for automatic analysis of audio including music and speech|
|US6549643||Nov 30, 1999||Apr 15, 2003||Siemens Corporate Research, Inc.||System and method for selecting key-frames of video data|
|US6564263 *||Dec 3, 1999||May 13, 2003||International Business Machines Corporation||Multimedia content description framework|
|US6567980 *||Aug 14, 1998||May 20, 2003||Virage, Inc.||Video cataloger system with hyperlinked output|
|US6626954 *||Feb 9, 1999||Sep 30, 2003||Sony Corporation||Information processing apparatus/method and presentation medium|
|US6628303 *||Jul 29, 1996||Sep 30, 2003||Avid Technology, Inc.||Graphical user interface for a motion video planning and editing system for a computer|
|US6665318 *||May 13, 1999||Dec 16, 2003||Hitachi, Ltd.||Stream decoder|
|US6686970||Oct 5, 1998||Feb 3, 2004||Canon Kabushiki Kaisha||Multi-media editing method and apparatus|
|US6697523||Aug 9, 2000||Feb 24, 2004||Mitsubishi Electric Research Laboratories, Inc.||Method for summarizing a video using motion and color descriptors|
|US6697564 *||Mar 3, 2000||Feb 24, 2004||Siemens Corporate Research, Inc.||Method and system for video browsing and editing by employing audio|
|US6715126 *||Sep 15, 1999||Mar 30, 2004||International Business Machines Corporation||Efficient streaming of synchronized web content from multiple sources|
|US6721454 *||Jul 1, 1999||Apr 13, 2004||Sharp Laboratories Of America, Inc.||Method for automatic extraction of semantically significant events from video|
|US6757027||Sep 27, 2000||Jun 29, 2004||Sony Corporation||Automatic video editing|
|US6760916 *||Apr 18, 2001||Jul 6, 2004||Parkervision, Inc.||Method, system and computer program product for producing and distributing enhanced media downstreams|
|US6774917 *||Mar 11, 1999||Aug 10, 2004||Fuji Xerox Co., Ltd.||Methods and apparatuses for interactive similarity searching, retrieval, and browsing of video|
|US6873740 *||Nov 28, 2000||Mar 29, 2005||Koninklijke Philips Electronics N.V.||Method for coding and decoding multimedia data|
|US6877134 *||Jul 29, 1999||Apr 5, 2005||Virage, Inc.||Integrated data and real-time metadata capture system and method|
|US6898291 *||Jun 30, 2004||May 24, 2005||David A. Gibson||Method and apparatus for using visual images to mix sound|
|US6898759 *||Nov 20, 1998||May 24, 2005||Yamaha Corporation||System of generating motion picture responsive to music|
|US6904227 *||Feb 15, 2000||Jun 7, 2005||Nec Corporation||Device and method for editing video and/or audio data recorded in a disc storage medium|
|US6912327 *||Jan 28, 2000||Jun 28, 2005||Kabushiki Kaisha Toshiba||Imagine information describing method, video retrieval method, video reproducing method, and video reproducing apparatus|
|US6912726 *||Apr 2, 1997||Jun 28, 2005||International Business Machines Corporation||Method and apparatus for integrating hyperlinks in video|
|US6934906 *||Jul 10, 2000||Aug 23, 2005||At&T Corp.||Methods and apparatus for integrating external applications into an MPEG-4 scene|
|US7020888 *||Nov 27, 2001||Mar 28, 2006||Intellocity Usa, Inc.||System and method for providing an omnimedia package|
|US7024677 *||Aug 8, 2000||Apr 4, 2006||Thomson Licensing||System and method for real time video production and multicasting|
|US7093191 *||Dec 21, 2001||Aug 15, 2006||Virage, Inc.||Video cataloger system with synchronized encoders|
|US7356830 *||Jul 9, 1999||Apr 8, 2008||Koninklijke Philips Electronics N.V.||Method and apparatus for linking a video segment to another segment or information source|
|US7362946 *||Apr 5, 2000||Apr 22, 2008||Canon Kabushiki Kaisha||Automated visual image editing system|
|US7702014 *||Dec 16, 1999||Apr 20, 2010||Muvee Technologies Pte. Ltd.||System and method for video production|
|US7725812 *||Mar 31, 2000||May 25, 2010||Avid Technology, Inc.||Authoring system for combining temporal and nontemporal digital media|
|US20010009423 *||Mar 28, 2001||Jul 26, 2001||Marc Davis||Time-based media processing system|
|US20010012444 *||Feb 9, 2001||Aug 9, 2001||Masamichi Ito||Image processing method and apparatus|
|US20020032697 *||Jun 6, 2001||Mar 14, 2002||Synapix, Inc.||Time inheritance scene graph for representation of media content|
|US20020053078 *||Apr 18, 2001||May 2, 2002||Alex Holtz||Method, system and computer program product for producing and distributing enhanced media downstreams|
|US20020069218 *||Jul 23, 2001||Jun 6, 2002||Sanghoon Sull||System and method for indexing, searching, identifying, and editing portions of electronic multimedia files|
|US20030233929 *||Jun 20, 2002||Dec 25, 2003||Koninklijke Philips Electronics N.V.||System and method for indexing and summarizing music videos|
|US20040125124 *||Feb 18, 2003||Jul 1, 2004||Hyeokman Kim||Techniques for constructing and browsing a hierarchical video structure|
|US20060029368 *||Oct 12, 2005||Feb 9, 2006||Vulcan Patents, Llc||Iterative, maximally probable, batch-mode commercial detection for audiovisual content|
|US20080016114 *||Jul 14, 2006||Jan 17, 2008||Gerald Thomas Beauregard||Creating a new music video by intercutting user-supplied visual data with a pre-existing music video|
|AU2002521A||Title not available|
|AU200025214B2||Title not available|
|EP1058453A1||Dec 14, 1999||Dec 6, 2000||Matsushita Electric Industrial Co., Ltd.||Method and apparatus for retrieving desired video and audio scene using voice recognition|
|JP2000268540A||Title not available|
|JPH0668168A||Title not available|
|WO2000039997A2||Dec 29, 1999||Jul 6, 2000||Earthnoise.Com Inc.||Creating and editing digital video movies|
|1||Ahanger, G. and T.D.C. Little, Automatic Composition Techniques for Video Production.|
|2||Chua, Tat-Seng and Li-Qun Ruan, "A Video Retrieval and Sequencing System".|
|3||Effelsberg, Prof. Dr. W., "Automatische Inhaltsanalyse von digitalen Videos, 4.5.6".|
|4||http://vaio-online.sony.com/sg/vaio/Press/pr4-content.html; VAIO Online Singapore, What's New, "Press Release"; Oct. 24, 2000.|
|5||http://vaio-online.sony.com/sg/vaio/Press/pr4—content.html; VAIO Online Singapore, What's New, "Press Release"; Oct. 24, 2000.|
|6||Lienhart, Rainer, Silvia Pfeiffer and Wolfgang Effelsberg, "Scene Determination based on Video and Audio Features".|
|7||Liu, Zhu, Yao Wang and Tsuhan Chen, "Audio Feature Extraction and Analysis for Scene Segmentation and Classification".|
|8||Nack, Frank-Michael, Auteur: "The Application of Video Semantics and Theme Representation for Automated Film Editing", Aug. 1996.|
|9||Parkes, Alan P., "Temporal Viewpoints for Video-Based Question Answering", 1993 (presented as Culture, Technology, Interpretation: the Challenge of Multimedia, Trinity College, Dublin, Ireland).|
|10||Pfeiffer, S., R. Lienhart, S. Fischer and W. Effelsberg, "Abstracting Digital Movies Automatically".|
|11||Rowe, Robert and Eric L. Singer, "Two Highly Integrated Real-Time Music and Graphics Performance Systems".|
|12||Takahasi, Shogo, Kenji Suzuki, Hideyuki Sawada and Shuji Hashimoto, "Music Creation from Moving Image and Environmental Sound".|
|13||Tsekeridou, Sofia and Ioannis Pitas, "Audio-Visual Content Analysis for Content-Based Video Indexing".|
|Citing Patent||Filing date||Publication date||Applicant||Title|
|US8126190||Jan 31, 2007||Feb 28, 2012||The Invention Science Fund I, Llc||Targeted obstrufication of an image|
|US8126938||May 25, 2007||Feb 28, 2012||The Invention Science Fund I, Llc||Group content substitution in media works|
|US8203609||Jan 31, 2007||Jun 19, 2012||The Invention Science Fund I, Llc||Anonymization pursuant to a broadcasted policy|
|US8347210 *||Sep 21, 2009||Jan 1, 2013||Apple Inc.||Synchronizing video with audio beats|
|US8347212 *||Sep 20, 2010||Jan 1, 2013||Lifereel, Inc.||Presentation production system with universal format|
|US8357847 *||Jul 12, 2007||Jan 22, 2013||Mxp4||Method and device for the automatic or semi-automatic composition of multimedia sequence|
|US8631326 *||Aug 31, 2009||Jan 14, 2014||Apple Inc.||Segmented timeline for a media-editing application|
|US8661096 *||Nov 5, 2007||Feb 25, 2014||Cyberlink Corp.||Collaborative editing in a video editing system|
|US8698877 *||Apr 22, 2010||Apr 15, 2014||Sony Corporation||Image information processing apparatus, image capture apparatus, image information processing method, and program|
|US8732087||Mar 30, 2007||May 20, 2014||The Invention Science Fund I, Llc||Authorization for media content alteration|
|US8769421||Aug 31, 2009||Jul 1, 2014||Apple Inc.||Graphical user interface for a media-editing application with a segmented timeline|
|US8792673||Aug 5, 2011||Jul 29, 2014||The Invention Science Fund I, Llc||Modifying restricted images|
|US8805744 *||Jul 18, 2008||Aug 12, 2014||Disney Enterprises, Inc.||Podblasting-connecting a USB portable media device to a console|
|US8910033||May 25, 2007||Dec 9, 2014||The Invention Science Fund I, Llc||Implementing group content substitution in media works|
|US9065979||Sep 19, 2007||Jun 23, 2015||The Invention Science Fund I, Llc||Promotional placement in media works|
|US9092928||Aug 30, 2007||Jul 28, 2015||The Invention Science Fund I, Llc||Implementing group content substitution in media works|
|US9215512||Jun 6, 2011||Dec 15, 2015||Invention Science Fund I, Llc||Implementation of media content alteration|
|US9230601||Nov 25, 2008||Jan 5, 2016||Invention Science Fund I, Llc||Media markup system for content alteration in derivative works|
|US9286383||Aug 28, 2014||Mar 15, 2016||Sonic Bloom, LLC||System and method for synchronization of data and audio|
|US9343053||May 7, 2014||May 17, 2016||Sound In Motion||Adding audio sound effects to movies|
|US9426387||Jan 31, 2007||Aug 23, 2016||Invention Science Fund I, Llc||Image anonymization|
|US9495608 *||Jul 2, 2013||Nov 15, 2016||Sony Corporation||Smart feeling sensing tag for pictures|
|US9583141||May 28, 2008||Feb 28, 2017||Invention Science Fund I, Llc||Implementing audio substitution options in media works|
|US9781377||Mar 25, 2014||Oct 3, 2017||Tivo Solutions Inc.||Recording and playback system based on multimedia content fingerprints|
|US20070299877 *||May 25, 2007||Dec 27, 2007||Searete Llc||Group content substitution in media works|
|US20090088255 *||Jul 18, 2008||Apr 2, 2009||Disney Enterprises, Inc., A Delaware Corporation||Podblasting-connecting a usb portable media device to a console|
|US20090119369 *||Nov 5, 2007||May 7, 2009||Cyberlink Corp.||Collaborative editing in a video editing system|
|US20100050854 *||Jul 12, 2007||Mar 4, 2010||Mxp4||Method and device for the automatic or semi-automatic composition of multimedia sequence|
|US20100080532 *||Sep 21, 2009||Apr 1, 2010||Apple Inc.||Synchronizing Video with Audio Beats|
|US20100088602 *||Oct 3, 2008||Apr 8, 2010||Microsoft Corporation||Multi-Application Control|
|US20100266262 *||May 5, 2010||Oct 21, 2010||Samsung Electronics Co., Ltd.||Information storage medium containing subtitle data for multiple languages using text data and downloadable fonts and apparatus therefor|
|US20100266265 *||May 5, 2010||Oct 21, 2010||Samsung Electronics Co., Ltd.||Information storage medium containing subtitle data for multiple languages using text data and downloadable fonts and apparatus therefor|
|US20100281381 *||Aug 31, 2009||Nov 4, 2010||Brian Meaney||Graphical User Interface for a Media-Editing Application With a Segmented Timeline|
|US20100281383 *||Aug 31, 2009||Nov 4, 2010||Brian Meaney||Segmented Timeline for a Media-Editing Application|
|US20110071931 *||Sep 20, 2010||Mar 24, 2011||Negley Mark S||Presentation Production System With Universal Format|
|US20110214045 *||Jan 31, 2011||Sep 1, 2011||Jason Sumler||System, method, and computer readable medium for creating a video clip|
|US20110261160 *||Apr 22, 2010||Oct 27, 2011||Sony Corporation||Image information processing apparatus, image capture apparatus, image information processing method, and program|
|US20120154679 *||Dec 21, 2010||Jun 21, 2012||Comcast Cable Communications, Llc||User-controlled synchronization of audio and video|
|US20150356120 *||Nov 14, 2014||Dec 10, 2015||Fuji Xerox Co., Ltd.||Design management apparatus, design management method, and non-transitory computer readable medium|
|U.S. Classification||715/728, 715/721, 715/726, 715/723|
|International Classification||G11B27/034, H04N5/91, G06F3/16, G11B27/34|
|Cooperative Classification||G11B27/034, G11B27/34|
|European Classification||G11B27/34, G11B27/034|
|Jun 18, 2003||AS||Assignment|
Owner name: MUVEE TECHNOLOGIES PTE LTD., SINGAPORE
Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNORS:KELLOCK, PETER ROWAN;ALTMAN, EDWARD JAMES;REEL/FRAME:014437/0026;SIGNING DATES FROM 20030605 TO 20030609
Owner name: MUVEE TECHNOLOGIES PTE LTD., SINGAPORE
Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNORS:KELLOCK, PETER ROWAN;ALTMAN, EDWARD JAMES;SIGNING DATES FROM 20030605 TO 20030609;REEL/FRAME:014437/0026
|Feb 11, 2015||FPAY||Fee payment|
Year of fee payment: 4